Research shows that 80% of all occupations will be affected by AI in the future, jobs affected by ChatGPT tend to have higher wages
Research has revealed that 80% of workers may be affected by the use of chat AI using a large-scale language model such as ``ChatGPT''. In addition, as a trend, wages are high, and it is believed that the impact on jobs that require programming and writing skills is large.
[2303.10130] GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models
OpenAI: ChatGPT Could Disrupt 19% of US Jobs, Is Yours on the List? | PCMag
OpenAI Research Says 80% of US Workers' Jobs Will Be Impacted by GPT
Researchers at OpenAI and the University of Pennsylvania, who are developing ChatGPT, investigated how much AI affects the automation of work and posted the results on the dissertation publishing site arXiv.
In the survey, 1016 occupations were listed using O*NET, a major occupational database in the United States, and tasks to be measured for each occupation were determined. Use a rubric to determine whether direct access to GPT or access to a secondary system using GPT can reduce the time required to perform a specific task by at least 50% compared to what a human would do Did.
According to this content, 80% of American workers are affected by 10% of the tasks they have, and 19% of them may be affected by 50% of the tasks.
Professions most likely to be affected include mathematicians, tax preparers, writers, web designers, accountants, journalists, legal secretaries, and interpreters/translators. On the other hand, occupations that are less likely to be affected include graphic designers, search marketing analysts, and financial managers.
By industry, data processing services, information services, publishing and insurance services are likely to be affected, while food manufacturing, wood product manufacturing and agricultural support activities will be less affected. Jobs with higher wages and jobs that require programming or writing skills tend to be more susceptible.
As mentioned above, in this research, judgment is made by ``whether the task that is the object of measurement can be shortened'', so whether the task can be completely decomposed in the first place, and the specific category necessary to carry out the task well Note that I am not sure if I am systematically omitting skills or tasks for
Since chat AI may lie like that, it is not considered that the content needs to be checked by humans, and even if the task can be shortened, it is not necessarily completely automated. It has also been pointed out that
“Large-scale language models have consistently improved their capabilities over time. Today, even if we stop developing new capabilities, the growing economic impact of large-scale language models will not persist.” We expect it to increase,” he said.
Related Posts:
in Note, Posted by logc_nt